linux-kernel - Re: make-bogus-warnings-go-away tree [was: 2.6.18-mm3]

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20061005163721.GJ16812@stusta.de>
Date:	Thu, 5 Oct 2006 18:37:21 +0200
From:	Adrian Bunk <bunk@...sta.de>
To:	Ingo Molnar <mingo@...e.hu>
Cc:	Jeff Garzik <jeff@...zik.org>, linux-kernel@...r.kernel.org,
	Andrew Morton <akpm@...l.org>
Subject: Re: make-bogus-warnings-go-away tree [was: 2.6.18-mm3]

On Thu, Oct 05, 2006 at 10:37:54AM +0200, Ingo Molnar wrote:
> 
> * Andrew Morton <akpm@...l.org> wrote:
> 
> > - Added Jeff's make-bogus-warnings-go-away tree to the -mm lineup, as
> >   git-gccbug.patch
> 
> Jeff: very nice! (I did this myself on a much smaller scale for the -rt 
> patch, because it's just so lethal if some serious warning gets lost in 
> the myriads of 'possible use of uninitialized' messages.)
> 
> A small suggestion: to give GCC folks a chance to actually fix this, 
> could we actively annotate these places instead of working them around?
> 
> I.e., instead of:
> 
>         long cursor = 0;
>         int error = 0;
> -       void *new_mc;
> +       void *new_mc = NULL;
>         int cpu;
>         cpumask_t old;
> 
> couldnt we do:
> 
> 	void *new_mc __GCC_WARN_BUG;
> 
> and then do something like this in gcc.h:
> 
>  #ifdef CONFIG_ELIMINATE_BOGUS_GCC_WARNINGS
>  # define __GCC_WARN_BUG = 0
>  #else
>  # define __GCC_WARN_BUG
>  #endif
> 
> this both gives an in-source incentive for GCC folks to get rid of these 
> bogus warnings (or remain shamed for eternity),

Not all of the false positives are gcc bugs.

There are cases where it's technically impossible for gcc to figure out 
that a variable is always initialized.

> and gives us the ability 
> to control the presence of these workarounds (and the eventual ability 
> to eliminate them in the future).
> 
> this would also mean we could merge your tree upstream without worrying 
> about hiding gcc bugs.

What we'd need would be some -Wno-may-be-used-uninitialized gcc option 
that turns off the "may be may be used uninitialized" warnings but not 
the "is used uninitialized" warnings.

This would:
- give us a way to silence these warnings
- allow people to see the warnings if they want to
- not increase the maintenance overhead

> 	Ingo

cu
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/